Longitudinal impact of cloud-based medical imaging: an analysis using ARIMA modeling and Monte Carlo simulation

Cloud-based medical imaging systems represent a paradigm shift in healthcare digitization. However, existing evaluations often rely on linear models that overlook critical methodological issues, particularly time-series autocorrelation in longitudinal data and uncertainty quantification in economic assessments.

To quantify the geographic, economic, and film-related environmental impacts of cloud medical imaging adoption using ARIMA modeling and Monte Carlo simulation.

We analyzed 15,709 cleaned cloud access records from November 2021 to March 2025 and 464 month-by-format film utilization records covering 114 calendar months from January 2015 to June 2024 at a major tertiary hospital. Time-series autocorrelation was evaluated using ARIMA modeling. Geographic inequality was assessed using the Gini coefficient. Economic robustness was evaluated using a six-scenario deterministic sensitivity analysis and Monte Carlo simulation (10,000 iterations).

The system exhibited marked geographic concentration (Gini coefficient = 0.980, 95% CI: 0.977-0.982), with 85.6% of access originating from local and provincial users. ARIMA(1,1,2) modeling of monthly film savings suggested an upward trend in film savings (monthly increase: 281.5 sheets). Forecasts suggested a plateauing trend rather than indefinite growth. Monte Carlo simulation yielded a mean estimated savings of ¥15.75 million (95% CI: [¥13.16, ¥18.35] million). Film-related avoided emissions were estimated at 341 tons of CO2e.

This single-center study suggests that cloud imaging can be associated with substantial film-related cost savings and avoided emissions in a large urban tertiary hospital. Broader generalizability, net system-wide economic benefit, and full environmental impact require multicenter validation and more comprehensive life-cycle assessment.